Journal of Molecular Biology
Regular articleStrategies for the in vitro evolution of protein function: enzyme evolution by random recombination of improved sequences1
Introduction
Enzymes can be evolved in vitro to exhibit new and useful functions. A sampling of the local sequence space of the enzyme is created by mutagenesis; screening or selection directs the evolution towards the desired features. A successful strategy for improving enzyme activity in non-natural environments (Chen & Arnold, 1993) and on non-natural substrates (Moore & Arnold, 1996) has been to accumulate amino acid substitutions over multiple generations of random mutagenesis and screening. In practice, the best variant identified in each generation is chosen to parent the subsequent generation. Other potentially useful variants are set aside, and their mutations must be rediscovered in the evolved protein background in order to become incorporated. Because there is no mechanism other than back mutation for deleting mutations, this approach can also accumulate deleterious mutations, leading to premature termination of an evolving lineage. These are the classical arguments for the benefits of recombination (sex) in evolution (Maynard Smith, 1988). Recombination allows more rapid accumulation of beneficial mutations present in a population. It also makes possible the removal of deleterious mutations which would otherwise accumulate in an asexual population, a phenomenon known to geneticists as Müller’s ratchet (Müller, 1932). Recombination can provide similar benefits for in vitro molecular evolution Stemmer 1994a, Stemmer 1994b.
Bacillus subtilis p-nitrobenzyl (pNB) esterase catalyzes the hydrolysis of the para-nitrobenzyl esters of various cephalosporin-type antibiotics, a necessary step in their large-scale synthesis (Zock et al., 1994). Using four generations of sequential random mutagenesis and screening, we evolved a series of pNB esterases up to 30 times more active towards hydrolysis of the pNB ester of loracarbef (LCN-pNB) in aqueous dimethylformamide (Moore & Arnold, 1996). During the fourth generation, a large number (∼7500) of pNB esterase clones were screened and partially characterized in order to validate the rapid screening assay. Sixteen improved pNB esterase clones were identified, from which the five most active enzymes (>50% enhancements in activity over the parent enzyme) were characterized. DNA sequencing revealed four unique pNB esterases (Table 1). Due to the limitations of screening, evolved sequences are generated using a low rate of point mutagenesis and typically accumulate a single beneficial mutation per generation. A simple restriction/ligation experiment demonstrated that recombination of mutations present in at least two of those sequences could further improve pNB esterase activity. Recombining gene segments from two improved pNB esterase variants yielded an enzyme twice as active as the best parent. DNA sequencing demonstrated that mutations from each of the two parents were combined in the new sequence (I60V and L334S), while one neutral or slightly deleterious mutation was deleted (K267R; Moore & Arnold, 1996).
Stemmer recently introduced the technique of “DNA shuffling” to create novel genes by recombination of closely-related DNA sequences (Stemmer, 1994b). Because it also introduces new point mutations during reassembly of the DNA fragments, DNA shuffling alone has been effective for directed protein evolution starting from a single sequence Stemmer 1994a, Crameri et al 1996. Questions arise as to how this approach is best implemented and integrated with other in vitro evolution approaches such as sequential random mutagenesis. Issues include optimizing the point mutagenesis rate associated with DNA shuffling, determining appropriate screening sample sizes and how many parental genes to recombine, and deciding when to use recombination. Here we investigate the further evolution of pNB esterase by DNA shuffling of the improved sequences generated by random mutagenesis and screening. By following how the genes evolve during cycles of DNA shuffling and screening, we can elucidate the mechanisms contributing to the evolution of function and begin to optimize strategies for in vitro evolution. An analysis of the recombination process identifies some of its benefits and limitations for directed evolution and allows a rational choice of mutagenesis and screening strategies.
Section snippets
Recombination statistics and screening requirements
To comment on the utility of DNA shuffling in directed evolution, a review of the statistics of recombination of multiple parent sequences is useful. For this discussion, we will assume that the mutations are unique and distributed far enough from one another on the genes that recombination occurs freely between any two. Furthermore, equal amounts of the initial DNA sequences are recombined. Consider the random recombination of three parent sequences, each of which contains a si ngle mutation.
Conclusions
Recombination is an important tool for directing the evolution of proteins. Beneficial mutations can be recombined, while neutral and deleterious mutations are eliminated. The need to screen rather than select for many important enzyme functions, however, severely limits the ability to search for useful combinations. It is therefore imperative to analyze various recombination strategies. Mutagenic rates associated with the recombination process must be low so that beneficial mutations are not
DNA shuffling
DNA shuffling was performed as described by Stemmer (1994b) with modifications. The 2 kb DNA fragment encoding the B. subtilis pNB esterase gene was amplified using PCR (forward primer 5′-CAATCTAGAGGGTATTAATAATG-3′ and reverse primer 5′-CGCGGGATCCCCGGGTACCGGGC-3′). The amplified DNA was purified by gel electrophoresis and extraction using Qiaex kit (Qiagen, Chatsworth, CA). A total quantity of ∼10 μg DNA, either from a single parent (non-recombinatorial) or from a mixture of multiple parent
Acknowledgements
The authors thank Dr W. P. C. Stemmer (Maxygen) for many helpful discussions, Dr Steve Queener (Eli Lilly & Co.) for providing us the wild-type pNB esterase and the challenge to evolve it, and Ms Rebecca Little (Eli Lilly & Co.) for assistance with DNA sequencing. This research is supported by the US Department of Energy’s program in Biological and Chemical Technologies Research within the Office of Industrial Technologies, Energy Efficiency and Renewables. O. K. is supported by an NIH
References (15)
- et al.
The Bacillus subtilis pnbA gene encoding p-nitrobenzyl esterasecloning, sequence and high-level expression in Escherichia coli
Gene
(1994) - et al.
Tuning the activity of an enzyme for unusual environmentssequential random mutagenesis of subtilisin E for catalysis in dimethylformamide
Proc. Natl Acad. Sci. USA
(1993) - et al.
Purification and properties of p-nitrobenzyl esterase from Bacillus subtilis
J. Ind. Micro.
(1995) - et al.
Improved green fluorescent protein by molecular evolution using DNA shuffling
Nature Biotech.
(1996) The Evolution of Recombination
- et al.
Directed evolution of a para-nitrobenzyl esterase for aqueous-organic solvents
Nature Biotech.
(1996) - et al.
Optimization of industrial enzymes by directed evolution
Advan. Biochem. Eng.
(1997)
Cited by (131)
Construction and analysis of Sip1Aa insecticidal protein random recombination library
2021, Journal of Asia-Pacific EntomologyCitation Excerpt :Random mutation is generally used in the studies of directed evolution (Cameron, 2004; Lutz and Patrick, 2004; Kaur, 2006), by error-prone PCR and chemical mutagenesis (Hida et al., 2007; Fujii et al., 2004). The error-prone PCR is widely used method easy to operate and has a high mutation rate (Fujii et al., 2004; Rasila et al., 2009; Moore et al., 1997). This approach is based on the high error rate inherent in the Taq enzyme, and the mutation rate can be adjusted by adding nucleotide analogues or changing the bivalent cation concentration in the reaction (Spee et al., 1993; Shafikhani et al., 1997).
Structure-function relationship of extremozymes
2021, Microbial Extremozymes: Novel Sources and Industrial ApplicationsEnhanced the catalytic efficiency and thermostability of maltooligosyltrehalose synthase from Arthrobacter ramosus by directed evolution
2020, Biochemical Engineering JournalCitation Excerpt :Directed evolution is a process that simulates the effects of natural evolution to allow screening of viable mutants from a large library. Given the size of these libraries, the probability of screening positive mutants is high, which is why it is widely used in protein engineering [22–24]. Zhang et al. increased the catalytic efficiency (kcat/Km) of d-psicose 3-epimerase from Clostridium bolteae by 20 %, with increasing the product concentration by 50 % by direct evolution [25].
Protein engineering approaches for enhanced catalytic efficiency
2020, Biomass, Biofuels, Biochemicals: Advances in Enzyme Catalysis and TechnologiesEnhancing the thermostability of fumarase C from Corynebacterium glutamicum via molecular modification
2018, Enzyme and Microbial TechnologyCitation Excerpt :The effective method is to engineer existing biocatalysts to match the target conditions of industrial production via directed evolution [20,21]. Considering that the error-prone PCR (ep PCR) is simple, rapid and inexpensive [22], it was uesd to introduce mutations into the target gene and construct the mutant library. The recombinant plasmid pET28a (+)-cgFumC were used as templates for ep PCR.
Remaining Challenges in the Metabolic Engineering of Yeasts for Biofuels
2015, Direct Microbial Conversion of Biomass to Advanced Biofuels
- 1
Edited by J. Wells
- 2
Present addresses: J. C. Moore, Merck & Co., Inc., P.O. Box 2000, RY80Y-110, Rahway, NJ 07065, USA; H.-M. Jin, SmithKline Beecham, Mail Code UE0548, 709 Swedeland Rd., King of Prussia, PA 19406, USA.